HOW YOUR
SCORE WORKS.

Am I Jacked rates visible muscular development in a single photo. The score is subjective entertainment. It is not a strength test, health assessment, body-fat estimate, attractiveness rating or population percentile.

Age and the scoring scale

The launch experience is for adults 18 and older, including photo scoring, normal assessment feedback and optional Savage roasts, body leaderboards, goal images and the paid training plan. Adult scores use the same absolute visible-muscle rubric; entering a different adult age does not change the scale. Your 18+ confirmation controls access, and photo screening can reject possible minors or unclear adulthood. An appearance estimate cannot overrule an under-18 declaration.

Six areas, one transparent calculation

Shoulders, chest, arms, back, core and legs each receive a score from 0 to 100 only when adequately visible. Covered or cropped areas receive “not visible,” never a guessed number. The overall score is the rounded, equally weighted mean of the visible scores. At least two areas are needed for a private score; four are required for leaderboard eligibility.

Area anchors: 0–19 little visible development; 20–39 modest; 40–59 noticeable; 60–79 pronounced; 80–94 very pronounced; 95–100 exceptional. Your overall title uses the same six bands, so every title answers the question in the name: Starting Line (0–19, not yet), Base Building (20–39, getting there), Clearly Lifts (40–59, almost), Jacked (60–79, yes), Seriously Jacked (80–94, very) and Final Boss (95–100). Savage mode swaps in a blunter title for the same band, such as “Fine. You’re jacked.”; the public leaderboard always shows the standard title. A build label is separate from rank.

Size and definition are separate

Definition describes visible muscle separation: limited, some, clear or sharp. It receives no numeric body-fat estimate and never changes your muscle score or leaderboard position. When lighting or coverage prevents a fair call, it is unassessed. Visual build labels combine the muscle band with definition across the visible areas; they do not classify weight or health. Normal assessment labels use plain rubric words such as “Muscular · Smooth”. Savage labels such as “Offseason Mode” or “Factory Settings” are optional jokes, not claims about your diet, gender or body-fat percentage. Older results without definition keep their original findings.

Savage mode cannot move the score

The scoring call uses the same rubric whether roasting is enabled or not. The normal assessment includes your neutral feedback. A separate roast is generated only when you select Savage, using the same findings and fixed scores. Switching feedback modes does not reassess your photo. Opting into Savage mode adds a brutally honest RoastMe-style verdict, using a separate AI follow-up that preserves the original photo conversation, fixed scores, visible observations and photo limitations. Strong development earns credit. One sharp targeted roast is requested when a visible area scores below 60 and photo confidence is medium or high. A high overall score does not hide a real gap; hidden areas and low-confidence results are never jab targets. If there is no fair target, the verdict uses sarcastic respect or the actual photo limitation without inventing a weakness. It cannot change your score. Gym-meme titles describe score bands, not verified body types. We do not provide a natty score: a photo cannot establish steroid use.

Badges have actual thresholds

Boulder Shoulders, Sleeve Destroyer, Chest Day CEO, Lat Spread Lore and Core Main Character require at least 80 in the corresponding area. Leg Day Loyalist requires 75 in legs. No Muscle Left Behind requires at least four areas at 60 or above. Otherwise you get Origin Story. Hidden areas cannot earn a muscle badge.

Ranking consistency

The model uses the fixed written rubric. No reference pictures are used. Results can vary between runs and with lighting, clothing and pose. The scoring cohort is versioned. GPT-6.1 Sol uses visible-muscle-v2, keeping the written bands but separating it from older model rankings. Progress comparisons must use the same cohort; older saved results are not relabeled. Repeatability and subgroup performance have not been benchmarked against human judges.

A fairer photo

• Use the same natural light, camera distance and relaxed pose when comparing photos.
• Avoid filters, heavy pump lighting and altered images. They can affect visual scoring.
• Show your physique clearly and keep intimate areas fully covered. Shirtless photos are welcome; hidden muscles are not scored.
• All bodies use the same development rubric; gender, skin tone, health and attractiveness are not score inputs.

What a goal look means

Sol selects up to four visible areas and writes specific instructions for an optimistic, visibly stronger goal illustration. The changes should be noticeable while preserving believable anatomy and your original body fat and natural softness. The image model receives those instructions and a preservation prompt for your identity, clothing, pose and setting. Hidden anatomy stays unchanged; low-confidence results can reuse the original photo. It is an illustration. It offers no promise of what training will achieve, and no timeline. It cannot be submitted as your original leaderboard photo.

The leaderboard

Only server-produced scores from the current rubric can be posted. Each signed-in account has one entry; a new post replaces the old one. Ties sort by the original result time and then result ID. The board shows the top 50, with an optional rolling seven-day view. Photos, public participation and publication of your Savage roast require separate choices. The roast stays private unless you explicitly include it on your card or public profile. Reports can hide entries for review. AI detection is imperfect; this is friendly competition, not verified athletic judging.

Leaderboard photo screening

Before submission, a separate AI check looks for visible signs of synthetic generation or physique editing. Suspicious submissions are held for owner review. No visible flags does not prove a photo is authentic; realistic AI photos and subtle edits can pass, and genuine photos can be flagged. This screen does not change muscle scores.